Healthcare diagnosis decision support systems or computer-aided diagnosis or computer-assisted diagnosis (CAD) systems are used in medicine to assist users like medical experts or physicians in the interpretation of medical images. Imaging techniques in X-ray, magnetic resonance imaging (MRI), and Ultrasound diagnostics yield a great deal of information, which the user has to analyze and evaluate comprehensively in a short time. CAD systems help scan digital images, e.g. from magnetic resonance imaging, for typical appearances and to highlight conspicuous structures, such as vessels, nerve pathways, ventricles, functionally eloquent regions and/or tumor regions. Usually, machine-learning technologies, such as a decision tree and neural network, are utilized to build classifiers based on a large number of known cases with ground truth, i.e., cases for which the diagnosis has been confirmed by pathology. The classifier bases its diagnosis on a computational structure built from known cases and inputted features for the unknown structure case. The classifier output indicates the estimated nature of the unknown structure and optionally a confidence value. As the precision of medical imaging facilities improves to detect very small structures, and as the number of digital images to be processed increases this type of CAD becomes increasingly important as a tool to assist users like physicians. The computer-produced classification is considered a second opinion to a user like a physician in order to raise the accuracy and confidence associated with diagnosis.
Computer assisted surgery (CAS) represents a surgical concept and set of methods, that use computer technology for presurgical planning, and for guiding or performing surgical interventions. CAS is also known as computer aided surgery, computer assisted intervention, image guided surgery and surgical navigation, but these terms that are more or less synonyms with CAS.
The traditional approach of determining an applicable path of movement for a surgical and/or diagnostical device (a safe surgical trajectory) in an image guided therapy like MRI always comprises the two basic steps of: segmenting each critical structure in a spatial region defining the possible path(s) of movement around these regions and afterwards determining a corresponding safe or applicable path. One important area of application is brain surgery. In detail, the traditional surgery planning mainly follows the following steps:
In a first step, the target location is defined manually, or automatically, or in a semi-automated manner. This either involves registering the magnetic resonance (MR) volume to a template, often in the stereotactic coordinate system, and detection of anatomical structures from this transformation, or identifying some point and plane landmarks, such as Mid-sagittal plane, and AC/PC points to determine the location of the target. Once the target is determined, the planning reduces to the detection of an entry point. In many cases, the path between the entry and target points has to be straight and should not hit the critical structures.
The second step is the identification of critical structures, such as vessels, gyri and sulci, ventricles, and some functionally eloquent regions in the brain. This involves segmenting these structures for example from contrast-enhanced T1 weighted MR images by applying a set of image segmentation algorithms. Functionally eloquent regions are determined by analyzing fMRI (functional Magnetic Resonance images) and/or DTI (diffusion tensor images) data.